Eur J Ageing (2007) 4:145–154 DOI 10.1007/s10433-007-0054-4
ORIGINAL INVESTIGATION
Health and social service use among old people in the last 2 years of life Leena Forma Æ Pekka Rissanen Æ Anja Noro Æ Jani Raitanen Æ Marja Jylha¨
Published online: 8 August 2007 Ó Springer-Verlag 2007
Abstract This study focuses on differences in health and social service use in the last 2 years of life among Finnish people aged 70–79, 80–89, and 90 or older and on the variation in service use in the various municipalities. The data set, derived from multiple national registers, consists of 75,578 people who died in 1998–2001. The services included hospitals and long-term-care facilities, use of regular home care, and prescribed medicines. General hospital and public long-term care were the services most commonly used: general hospitals for younger age groups and public long-term care for older groups. The number of inpatient days in hospital was lower with increasing age, but older age groups used long-term care more frequently. Men had more hospital inpatient days than women, but women used more long-term care. The number of hospital inpatient days increased rapidly in the last months of life, almost doubling in the final month. Days in public longterm care increased regularly in the last 2 years of life. Variation in both hospital and long-term care by municipality was remarkable. The results indicate that, among people aged 70 years and older, age is a major determinant of care in the last 2 years of life. The variation in the use of care by municipality and the differences between men and women deserve more detailed analysis in future.
L. Forma (&) P. Rissanen J. Raitanen M. Jylha¨ Tampere School of Public Health, University of Tampere, Tampere 33014, Finland e-mail: [email protected] A. Noro National Research and Development Centre for Welfare and Health (STAKES), Helsinki, Finland
Keywords Last years of life Health and social care Use of care Ageing Older people Register study
Introduction There is a general understanding that people use more health and social care services in the last months and years of their lives than earlier, and that old people use more care than the young and middle-aged. Results, however, vary from one study to another by the type of care, the length of the period considered before death, and the time of the study. Also, the use of and demand for care preceding death is changing with practices in health and social care, and with increasing longevity in various populations. In this study, we examine the use of health and social care among old people in the last 2 years of their life in Finland, using register data covering both institutional and outpatient health and social services for the whole population. In particular, the focus is on the differences between age groups in old age. At the beginning of the Twenty-first century, about 70% of all deaths in Finland and 67% in the US occurred among persons aged 70 years or over (Statistics Finland 2006, personal notification; Kochanek et al. 2004). The proportion is expected to increase as a result of the population ageing and of the rapidly increasing life expectancy for those of both younger and older ages in particular. Several studies indicate that both age and closeness to death influence the need for care in older people. The likelihood of co-morbidity and functional decline increases with age, and, thus, need for services differs by age group. However, epidemiological studies indicate that old people who are dying experience a steeper decline in functional status than do same-age survivors (Guralnik et al. 1991; Wolinsky
123
146
et al. 1996), and among older people, the oldest are more likely to experience a longer period of disability before death (Lunney et al. 2003). Also, functional decline before death differs by age, being greater with more advanced age at death (Guralnik et al. 1991). A limited number of studies have concentrated on old people’s use of services in their last years of life; most of these have focused on costs of services. The studies vary according to the services included in the analysis. Most studies have included hospitalisation (e.g. Brameld et al. 1998; Seshamani and Gray 2004; Wilson and Truman 2002), but many have considered physician’s visits and home care also (e.g. Lubitz and Prihoda 1984; Bird et al. 2002; Hoover et al. 2002), and some have included support services (e.g. Mukamel et al. 2002). The results indicate that the use and costs of acute health care are greater in younger old people than in the most elderly (Shugarman et al. 2004; Seshamani and Gray 2004), but these differences more or less disappear or are even reversed when long-term care as well is taken into account (Hoover et al. 2002; McGrail et al. 2000). The follow-up time in relevant studies has varied from the last 3 months to the last 24 years of life, but most often service utilisation has been followed for the last year. The results indicate that the effect of closeness to death appears mainly in the last 2 years of life. Health care expenditure has been reported to increase most rapidly in the last 6 or final 3 months of life (e.g. Yang et al. 2003; Zweifel et al. 1999). Mukamel et al. (2002) found that health care service use increased 7 months before death; the largest increase was in the final month. Few of the previous studies have examined both hospital care and outpatient care. Also, most studies have been based on either administrative (e.g. Bickel 1998; Bird et al. 2002; Gaumer and Stavins 1991) or survey data (e.g. Brock et al. 1996), relying on self-reports or reports by the next of kin, which may be susceptible to recall problems. In Finland, one of the major advantages in health services research is the availability of comprehensive national registers that are based on provision of both health and social services, and are considered reliable. In Finland, municipalities, 431 in number in the year 2006, are responsible for organising health and social services for their residents. The municipalities may provide services themselves or in co-operation with other municipalities, or they can purchase them from private providers or other municipalities. These services are funded mainly by taxes, but also partly by user fees. In addition to this public service provision, the private sector produces about one-fifth of all health and social services. The proportion of private services, however, varies according to the type of service, being largest for dentists and outpatient care physicians, and very small for hospital care (3.6% for all hospital days in Finland, according to Hein et al. 2005). National health
123
Eur J Ageing (2007) 4:145–154
insurance, which usually covers all people regularly living in Finland, provides partial reimbursement for the costs of private services, including prescribed medicines also. Medication is included in the user fees for institutional care. Primary health care is provided by municipalities at health centres that also have hospital beds and provide both acute and long-term care. Specialised health care is produced by the municipalities via district and central hospitals owned by 20 hospital districts that are joint organisations. These districts form five university hospital districts, providing the most specialised health care. Municipalities and hospital districts have wide autonomy in organising the services, and there are differences between municipalities with regard to practices and also provision of care. Local service patterns influence social and health care usage and, to some extent, are also likely to affect the impact of age and time to death on service use. In this study, the focus is on age differences in health and social service use in the last 2 years of life among people aged 70 years and older. The study is part of a more comprehensive project on ‘Costs of Care Towards the End of Life’ (COCTEL). The detailed research questions are: 1.
2.
3.
To what extent does use of different health and social services in the last 2 years of life differ between age groups 70–79, 80–89, and 90 years? Here, we are interested in both, the proportion of people who used different services and the service quantities they used. How does the use of different health and social services vary with time to death (starting from 24 months preceding death), and how is age associated with this variation? Are the possible age differences in social and health service use in the last 2 years of life maintained even if the variation in service organisations among municipalities is taken into account?
Methods Data The basic population of this study consists of all people who lived in Finland and died between 1998 and 2001 at the age of 70 years. The sample was identified from the Central Population Register (Statistics Finland). The study population consisted of two subgroups: 1. 2.
all those who died at the age of 70 or older in the year 1998 and a random sample (40%) of all people living in Finland who died in 1999–2001 at age 70 or older, taken on those who were alive at 31 December 1997.
Eur J Ageing (2007) 4:145–154
147
The total sample consisted of 75,578 decedents. Their service use was studied for the time period of 2 years before death (not two calendar years but 730 or 731 days until death). The data were combined from five registers, maintained by Statistics Finland, the National Research and Development Centre for Welfare and Health (STAKES), and the Social Insurance Institution of Finland (SII). All records in these registers contain the individual’s unique social security number, which makes it possible to link records within the register and to information in the other registers. The same personal ID is used for all social and health care purposes. The registers are introduced in brief here, and Table 1 describes the information from each that is used in this study. The Causes of Death Register of Statistics Finland contains basic demographic characteristics; dates and places of birth; and dates, causes, and circumstances of death.
Table 1 Registers and variables used Register
Variables
Statistics Finland Causes of death register
Age
Age of death
Gender
1 = man, 2 = woman
Date of death Municipality of residence STAKES Care register for health care
Type of hospital
University hospital Central hospital District hospital Private hospital Health centre (inpatient departments)
Date of admission Date of discharge Care register for social welfare
Type of institution
Residential home Housing with 24-h assistance for older people
Date of admission Date of discharge Home care census Clients of regular home care Social Insurance Institution Prescription database
Medicines by prescription
The Care Register for Health Care of STAKES covers all hospitals in Finland. It contains data on the provider of hospital services and on the patient, admission, discharge, diagnoses, and care received. The Care Register for Social Welfare, also maintained by STAKES, registers the care episodes of residents in all long-term-care institutions in Finland, since 1996. The register contains data on the provider of service, client, admission, and discharge to care, as well as on the services and care received. These two registers also include census information for those care episodes that continue beyond the end of each calendar year. The Home Care Census has been conducted since 1995; it is performed 1 day every second year in November. It includes clients of regular municipal home care, and services they have received in the previous month. The register contains information on the provider of service, client, admission, and discharge to care, and on the services and care received. The prescription database of the Social Insurance Institution covers prescribed medicines for which non-institutionalised people have claimed reimbursement from the SII. The database includes ATC code (Anatomical Therapeutic Chemical classification system), date of purchase, costs, and the SII reimbursement received. The prescription database covers 97% of all purchased prescribed outpatient medicines for which reimbursement has been provided (Klaukka 2004). The ethics committee of the Pirkanmaa hospital district discussed the research plan and concluded that they did not object to the research being undertaken on ethical grounds. Discrepancies were found for 0.2% of admissions; there were inpatient days after the date of death or the same admission was recorded twice. These admissions were removed from the data. Corrections were made to 0.15% of admissions. Most of them concerned admission dates that referred to the same admission but differed between the census and discharge data; in these cases, the admission date from the discharge data was replaced by the admission date from the census data.
ATC code Date of medicine purchase
Measures In this study, the data include (1) inpatient days in hospital, (2) days in long-term care, (3) regular home care (at least once a week), and (4) use of prescribed medicines. Hospitals include university hospitals, general hospitals (including central, district, and private hospitals), and inpatient departments of health centres when the patient remains in care for less than 90 days. Long-term care includes care in residential homes, housing with 24-h assistance for older people, and inpatient departments of health centres in cases where the patient stays for 90 days
123
148
Eur J Ageing (2007) 4:145–154
Chi-square tests were performed to test differences between age groups in the proportions of subjects who had used services. Since the distributions of all study variables were strongly skewed or bimodal, medians and upper and lower quartiles are presented instead of means and standard deviations. Kruskall–Wallis tests were performed to find out if there are differences between the three age groups in the quantity of services used, and Mann–Whitney U tests were done for analysing more exactly, which age groups differ in respect of the quantity of services used. All analyses were performed with the SPSS statistical software package (version 12.0.1).
or more. Long-term care is divided into public and private care; all three categories are included in public care, but only residential homes and housing with 24-h assistance for older people are provided privately also. Home care includes both home nursing and home help. Service use was analysed for age groups 70–79, 80–89, and 90 years (age at death), separately for men and women. Data were analysed also in 5-year age bands but are reported in 10-year strata because the results were quite similar in the two classification schemes. We present, first, the proportion of subjects who had used various services and, second, the quantity of services used (care days or number of medicines), with the latter calculated for only those who had used the service in question. Use of medicines is reported as the number of different prescribed medicines (ATC codes to an accuracy of seven characters; pharmaceutical ingredient) purchased. Home care is described only as the proportion that received regular service; the number of home care visits in the last 2 years of life was not available. The service use was calculated for 24 months before death in each age group and for men and women separately. The months were calculated individually as 24 times 30 or 31 days before death. The last month was deemed to end on the day of death. To study the influence of municipal differences, the services were analysed in two categories: hospital care and long-term care. One very small municipality had only one person in the data. In one municipality (184 persons in these data), services were organised differently from those in all other municipalities, definitions of hospital and longterm care were different, and there were considerably more hospital inpatient days than in other municipalities. These two municipalities were excluded from these analyses.
Results Descriptive In the sample there were 30,786 (40.0%) men and 44,792 (60.0%) women. The average age at death was 82.1 years, 80.0 for men and 83.5 for women. Age distribution by gender is shown in Table 2. The age and gender distribution of the sample followed the distribution of all deaths in the age groups 70 in Finland in those years, and the distribution did not vary according to the year when people had died. This was also true separately for those deaths occurring in the 40% sample of people in 1999–2001. The proportion of service users in different age groups Service use in the last 2 years of life varied considerably between different groups. Of all subjects, 4.7% (1.4–8.8% in different age and gender groups) did not use inpatient days at all in their last 2 years of life, while 13.9% spent at least their two final years of life in some institution. Thus, the minimum number of inpatient days was 0 and the maximum was 731. The most frequent place of care was a general hospital for men aged 70–89 and women 70– 79 years old and was public long-term care for men aged 90 and women aged 80 years (see Table 3). Use of university and general hospitals was less frequent for older age groups, but long-term care increased with older age
Analysis To ensure that the sample is representative of the basic population, its age and gender distributions were compared to those for all deaths at the age of 70 and over in the study years 1998–2001 in Finland, using the data from Statistics Finland. Table 2 Sample characteristics (n = 75,578) Age group
Study sample
All deceased 1998–2001
Men n
Women %
n
Men %
n
Women %
n
%
70–79
15,591
50.6
13,398
29.9
28,345
50.6
24,118
29.5
80–89
12,348
40.1
22,242
49.7
22,389
40.0
40,591
49.7
90
2,847
9.3
9,152
20.4
5,286
9.4
16,982
20.8
Total
30,786
100.0
44,792
100.0
56,020
100.0
81,691
100.0
123
Eur J Ageing (2007) 4:145–154
149
Table 3 Any use of services and medicines, by age and gender, in the last 24 months of life Men 70–79
Women 80–89
90
70–79
80–89
90
Hospital inpatient care
86.5
85.9
79.1
85.7
78.8
65.8
University hospital
32.1
23.9
16.9
32.2
21.9
13.7
General hospital
65.1
62.7
51.7
62.4
54.2
39.5
Health centre
a
46.9
53.5
53.5
47.5
48.4
43.1
Long-term care
26.1
46.0
64.1
36.5
60.1
77.8
Public long-term care
23.6
40.8
57.0
33.8
54.8
70.7
Private long-term care Home care
4.0 11.2
8.3 18.9
10.6 21.8
5.0 16.7
9.2 22.2
11.7 18.6
Medicines
89.8
84.7
75.0
86.8
76.6
60.0
All results as percentages Results of v2 tests: Among both men and women the differences between age groups for each service type were statistically significant (P 0.001). The only exceptions were hospital inpatient care, where the difference between men of 70–79 and 80–89 years of age was not statistically significant, and health centres, where there was not a statistically significant difference between men 80–89 and 90 years old or women 70–79 and 80–89 years old a
ORIGINAL INVESTIGATION
Health and social service use among old people in the last 2 years of life Leena Forma Æ Pekka Rissanen Æ Anja Noro Æ Jani Raitanen Æ Marja Jylha¨
Published online: 8 August 2007 Ó Springer-Verlag 2007
Abstract This study focuses on differences in health and social service use in the last 2 years of life among Finnish people aged 70–79, 80–89, and 90 or older and on the variation in service use in the various municipalities. The data set, derived from multiple national registers, consists of 75,578 people who died in 1998–2001. The services included hospitals and long-term-care facilities, use of regular home care, and prescribed medicines. General hospital and public long-term care were the services most commonly used: general hospitals for younger age groups and public long-term care for older groups. The number of inpatient days in hospital was lower with increasing age, but older age groups used long-term care more frequently. Men had more hospital inpatient days than women, but women used more long-term care. The number of hospital inpatient days increased rapidly in the last months of life, almost doubling in the final month. Days in public longterm care increased regularly in the last 2 years of life. Variation in both hospital and long-term care by municipality was remarkable. The results indicate that, among people aged 70 years and older, age is a major determinant of care in the last 2 years of life. The variation in the use of care by municipality and the differences between men and women deserve more detailed analysis in future.
L. Forma (&) P. Rissanen J. Raitanen M. Jylha¨ Tampere School of Public Health, University of Tampere, Tampere 33014, Finland e-mail: [email protected] A. Noro National Research and Development Centre for Welfare and Health (STAKES), Helsinki, Finland
Keywords Last years of life Health and social care Use of care Ageing Older people Register study
Introduction There is a general understanding that people use more health and social care services in the last months and years of their lives than earlier, and that old people use more care than the young and middle-aged. Results, however, vary from one study to another by the type of care, the length of the period considered before death, and the time of the study. Also, the use of and demand for care preceding death is changing with practices in health and social care, and with increasing longevity in various populations. In this study, we examine the use of health and social care among old people in the last 2 years of their life in Finland, using register data covering both institutional and outpatient health and social services for the whole population. In particular, the focus is on the differences between age groups in old age. At the beginning of the Twenty-first century, about 70% of all deaths in Finland and 67% in the US occurred among persons aged 70 years or over (Statistics Finland 2006, personal notification; Kochanek et al. 2004). The proportion is expected to increase as a result of the population ageing and of the rapidly increasing life expectancy for those of both younger and older ages in particular. Several studies indicate that both age and closeness to death influence the need for care in older people. The likelihood of co-morbidity and functional decline increases with age, and, thus, need for services differs by age group. However, epidemiological studies indicate that old people who are dying experience a steeper decline in functional status than do same-age survivors (Guralnik et al. 1991; Wolinsky
123
146
et al. 1996), and among older people, the oldest are more likely to experience a longer period of disability before death (Lunney et al. 2003). Also, functional decline before death differs by age, being greater with more advanced age at death (Guralnik et al. 1991). A limited number of studies have concentrated on old people’s use of services in their last years of life; most of these have focused on costs of services. The studies vary according to the services included in the analysis. Most studies have included hospitalisation (e.g. Brameld et al. 1998; Seshamani and Gray 2004; Wilson and Truman 2002), but many have considered physician’s visits and home care also (e.g. Lubitz and Prihoda 1984; Bird et al. 2002; Hoover et al. 2002), and some have included support services (e.g. Mukamel et al. 2002). The results indicate that the use and costs of acute health care are greater in younger old people than in the most elderly (Shugarman et al. 2004; Seshamani and Gray 2004), but these differences more or less disappear or are even reversed when long-term care as well is taken into account (Hoover et al. 2002; McGrail et al. 2000). The follow-up time in relevant studies has varied from the last 3 months to the last 24 years of life, but most often service utilisation has been followed for the last year. The results indicate that the effect of closeness to death appears mainly in the last 2 years of life. Health care expenditure has been reported to increase most rapidly in the last 6 or final 3 months of life (e.g. Yang et al. 2003; Zweifel et al. 1999). Mukamel et al. (2002) found that health care service use increased 7 months before death; the largest increase was in the final month. Few of the previous studies have examined both hospital care and outpatient care. Also, most studies have been based on either administrative (e.g. Bickel 1998; Bird et al. 2002; Gaumer and Stavins 1991) or survey data (e.g. Brock et al. 1996), relying on self-reports or reports by the next of kin, which may be susceptible to recall problems. In Finland, one of the major advantages in health services research is the availability of comprehensive national registers that are based on provision of both health and social services, and are considered reliable. In Finland, municipalities, 431 in number in the year 2006, are responsible for organising health and social services for their residents. The municipalities may provide services themselves or in co-operation with other municipalities, or they can purchase them from private providers or other municipalities. These services are funded mainly by taxes, but also partly by user fees. In addition to this public service provision, the private sector produces about one-fifth of all health and social services. The proportion of private services, however, varies according to the type of service, being largest for dentists and outpatient care physicians, and very small for hospital care (3.6% for all hospital days in Finland, according to Hein et al. 2005). National health
123
Eur J Ageing (2007) 4:145–154
insurance, which usually covers all people regularly living in Finland, provides partial reimbursement for the costs of private services, including prescribed medicines also. Medication is included in the user fees for institutional care. Primary health care is provided by municipalities at health centres that also have hospital beds and provide both acute and long-term care. Specialised health care is produced by the municipalities via district and central hospitals owned by 20 hospital districts that are joint organisations. These districts form five university hospital districts, providing the most specialised health care. Municipalities and hospital districts have wide autonomy in organising the services, and there are differences between municipalities with regard to practices and also provision of care. Local service patterns influence social and health care usage and, to some extent, are also likely to affect the impact of age and time to death on service use. In this study, the focus is on age differences in health and social service use in the last 2 years of life among people aged 70 years and older. The study is part of a more comprehensive project on ‘Costs of Care Towards the End of Life’ (COCTEL). The detailed research questions are: 1.
2.
3.
To what extent does use of different health and social services in the last 2 years of life differ between age groups 70–79, 80–89, and 90 years? Here, we are interested in both, the proportion of people who used different services and the service quantities they used. How does the use of different health and social services vary with time to death (starting from 24 months preceding death), and how is age associated with this variation? Are the possible age differences in social and health service use in the last 2 years of life maintained even if the variation in service organisations among municipalities is taken into account?
Methods Data The basic population of this study consists of all people who lived in Finland and died between 1998 and 2001 at the age of 70 years. The sample was identified from the Central Population Register (Statistics Finland). The study population consisted of two subgroups: 1. 2.
all those who died at the age of 70 or older in the year 1998 and a random sample (40%) of all people living in Finland who died in 1999–2001 at age 70 or older, taken on those who were alive at 31 December 1997.
Eur J Ageing (2007) 4:145–154
147
The total sample consisted of 75,578 decedents. Their service use was studied for the time period of 2 years before death (not two calendar years but 730 or 731 days until death). The data were combined from five registers, maintained by Statistics Finland, the National Research and Development Centre for Welfare and Health (STAKES), and the Social Insurance Institution of Finland (SII). All records in these registers contain the individual’s unique social security number, which makes it possible to link records within the register and to information in the other registers. The same personal ID is used for all social and health care purposes. The registers are introduced in brief here, and Table 1 describes the information from each that is used in this study. The Causes of Death Register of Statistics Finland contains basic demographic characteristics; dates and places of birth; and dates, causes, and circumstances of death.
Table 1 Registers and variables used Register
Variables
Statistics Finland Causes of death register
Age
Age of death
Gender
1 = man, 2 = woman
Date of death Municipality of residence STAKES Care register for health care
Type of hospital
University hospital Central hospital District hospital Private hospital Health centre (inpatient departments)
Date of admission Date of discharge Care register for social welfare
Type of institution
Residential home Housing with 24-h assistance for older people
Date of admission Date of discharge Home care census Clients of regular home care Social Insurance Institution Prescription database
Medicines by prescription
The Care Register for Health Care of STAKES covers all hospitals in Finland. It contains data on the provider of hospital services and on the patient, admission, discharge, diagnoses, and care received. The Care Register for Social Welfare, also maintained by STAKES, registers the care episodes of residents in all long-term-care institutions in Finland, since 1996. The register contains data on the provider of service, client, admission, and discharge to care, as well as on the services and care received. These two registers also include census information for those care episodes that continue beyond the end of each calendar year. The Home Care Census has been conducted since 1995; it is performed 1 day every second year in November. It includes clients of regular municipal home care, and services they have received in the previous month. The register contains information on the provider of service, client, admission, and discharge to care, and on the services and care received. The prescription database of the Social Insurance Institution covers prescribed medicines for which non-institutionalised people have claimed reimbursement from the SII. The database includes ATC code (Anatomical Therapeutic Chemical classification system), date of purchase, costs, and the SII reimbursement received. The prescription database covers 97% of all purchased prescribed outpatient medicines for which reimbursement has been provided (Klaukka 2004). The ethics committee of the Pirkanmaa hospital district discussed the research plan and concluded that they did not object to the research being undertaken on ethical grounds. Discrepancies were found for 0.2% of admissions; there were inpatient days after the date of death or the same admission was recorded twice. These admissions were removed from the data. Corrections were made to 0.15% of admissions. Most of them concerned admission dates that referred to the same admission but differed between the census and discharge data; in these cases, the admission date from the discharge data was replaced by the admission date from the census data.
ATC code Date of medicine purchase
Measures In this study, the data include (1) inpatient days in hospital, (2) days in long-term care, (3) regular home care (at least once a week), and (4) use of prescribed medicines. Hospitals include university hospitals, general hospitals (including central, district, and private hospitals), and inpatient departments of health centres when the patient remains in care for less than 90 days. Long-term care includes care in residential homes, housing with 24-h assistance for older people, and inpatient departments of health centres in cases where the patient stays for 90 days
123
148
Eur J Ageing (2007) 4:145–154
Chi-square tests were performed to test differences between age groups in the proportions of subjects who had used services. Since the distributions of all study variables were strongly skewed or bimodal, medians and upper and lower quartiles are presented instead of means and standard deviations. Kruskall–Wallis tests were performed to find out if there are differences between the three age groups in the quantity of services used, and Mann–Whitney U tests were done for analysing more exactly, which age groups differ in respect of the quantity of services used. All analyses were performed with the SPSS statistical software package (version 12.0.1).
or more. Long-term care is divided into public and private care; all three categories are included in public care, but only residential homes and housing with 24-h assistance for older people are provided privately also. Home care includes both home nursing and home help. Service use was analysed for age groups 70–79, 80–89, and 90 years (age at death), separately for men and women. Data were analysed also in 5-year age bands but are reported in 10-year strata because the results were quite similar in the two classification schemes. We present, first, the proportion of subjects who had used various services and, second, the quantity of services used (care days or number of medicines), with the latter calculated for only those who had used the service in question. Use of medicines is reported as the number of different prescribed medicines (ATC codes to an accuracy of seven characters; pharmaceutical ingredient) purchased. Home care is described only as the proportion that received regular service; the number of home care visits in the last 2 years of life was not available. The service use was calculated for 24 months before death in each age group and for men and women separately. The months were calculated individually as 24 times 30 or 31 days before death. The last month was deemed to end on the day of death. To study the influence of municipal differences, the services were analysed in two categories: hospital care and long-term care. One very small municipality had only one person in the data. In one municipality (184 persons in these data), services were organised differently from those in all other municipalities, definitions of hospital and longterm care were different, and there were considerably more hospital inpatient days than in other municipalities. These two municipalities were excluded from these analyses.
Results Descriptive In the sample there were 30,786 (40.0%) men and 44,792 (60.0%) women. The average age at death was 82.1 years, 80.0 for men and 83.5 for women. Age distribution by gender is shown in Table 2. The age and gender distribution of the sample followed the distribution of all deaths in the age groups 70 in Finland in those years, and the distribution did not vary according to the year when people had died. This was also true separately for those deaths occurring in the 40% sample of people in 1999–2001. The proportion of service users in different age groups Service use in the last 2 years of life varied considerably between different groups. Of all subjects, 4.7% (1.4–8.8% in different age and gender groups) did not use inpatient days at all in their last 2 years of life, while 13.9% spent at least their two final years of life in some institution. Thus, the minimum number of inpatient days was 0 and the maximum was 731. The most frequent place of care was a general hospital for men aged 70–89 and women 70– 79 years old and was public long-term care for men aged 90 and women aged 80 years (see Table 3). Use of university and general hospitals was less frequent for older age groups, but long-term care increased with older age
Analysis To ensure that the sample is representative of the basic population, its age and gender distributions were compared to those for all deaths at the age of 70 and over in the study years 1998–2001 in Finland, using the data from Statistics Finland. Table 2 Sample characteristics (n = 75,578) Age group
Study sample
All deceased 1998–2001
Men n
Women %
n
Men %
n
Women %
n
%
70–79
15,591
50.6
13,398
29.9
28,345
50.6
24,118
29.5
80–89
12,348
40.1
22,242
49.7
22,389
40.0
40,591
49.7
90
2,847
9.3
9,152
20.4
5,286
9.4
16,982
20.8
Total
30,786
100.0
44,792
100.0
56,020
100.0
81,691
100.0
123
Eur J Ageing (2007) 4:145–154
149
Table 3 Any use of services and medicines, by age and gender, in the last 24 months of life Men 70–79
Women 80–89
90
70–79
80–89
90
Hospital inpatient care
86.5
85.9
79.1
85.7
78.8
65.8
University hospital
32.1
23.9
16.9
32.2
21.9
13.7
General hospital
65.1
62.7
51.7
62.4
54.2
39.5
Health centre
a
46.9
53.5
53.5
47.5
48.4
43.1
Long-term care
26.1
46.0
64.1
36.5
60.1
77.8
Public long-term care
23.6
40.8
57.0
33.8
54.8
70.7
Private long-term care Home care
4.0 11.2
8.3 18.9
10.6 21.8
5.0 16.7
9.2 22.2
11.7 18.6
Medicines
89.8
84.7
75.0
86.8
76.6
60.0
All results as percentages Results of v2 tests: Among both men and women the differences between age groups for each service type were statistically significant (P 0.001). The only exceptions were hospital inpatient care, where the difference between men of 70–79 and 80–89 years of age was not statistically significant, and health centres, where there was not a statistically significant difference between men 80–89 and 90 years old or women 70–79 and 80–89 years old a
